In vivo bone strain experiments were performed on the ulnae of three female rhesus macaques to test how the bone deforms during locomotion. The null hypothesis was that, in an animal moving its limbs predominantly in sagittal planes, the ulna experiences anteroposterior bending. Three rosette strain gauges were attached around the circumference of the bone slightly distal to midshaft. They permit a complete characterization of the ulna's loading environment. Strains were recorded during walking and galloping activities. Principal strains and strain directions relative to the long axis of the bone were calculated for each gauge site. In all three animals, the lateral cortex experienced higher tensile than compressive principal strains during the stance phase of walking. Compressive strains predominated at the medial cortex of two animals (the gauge on this cortex of the third animal did not function). The posterior cortex was subject to lower strains; the nature of the strain was highly dependent on precise gauge position. The greater principal strains were aligned closely with the long axis of the bone in two animals, whereas they deviated up to 45 degrees from the long axis in the third animal. A gait change from walk to gallop was recorded for one animal. It was not accompanied by an incremental change in strain magnitudes. Strains are at the low end of the range of strain magnitudes recorded for walking gaits of nonprimate mammals. The measured distribution of strains in the rhesus monkey ulna indicates that mediolateral bending, rather than anteroposterior bending, is the predominant loading regime, with the neutral axis of bending running from anterior and slightly medial to posterior and slightly lateral. A variable degree of torsion was superimposed over this bending regime. Ulnar mediolateral bending is apparently caused by a ground reaction force vector that passes medial to the forearm. The macaque ulna is not reinforced in the plane of bending. The lack of buttressing in the loaded plane and the somewhat counterintuitive bending direction recommend caution with regard to conventional interpretations of long bone cross-sectional geometry.